Aprèn amb Pitavola: A Contemporary After-School Program For Awakening STEM Skills

Abstract

Over the last few years, the scope of the university, traditionally focused on young adults, has been revisited; universities have started an opening process towards today's information-driven society to contribute to the transformation of our education system. Since 2007, Universitat d'Andorra (UdA) has embraced this new era by implementing multiple initiatives in the framework of the UNESCO Chair Information Technologies: Training and solidary development. The case of small states. Among these initiatives, " Aprèn amb Pitavola " stands out as a contemporary approach to instil technology-related knowledge into primary schoolers. The main goal of this non-profit initiative is to bridge the gap between children's inherent ability to use technology and the actual interest in becoming more than mere users.

Full text

APRÈN AMB PITAVOLA: A CONTEMPORARY AFTER-SCHOOL
PROGRAM FOR AWAKENING STEM SKILLS
Aleix Dorca, Cristina Yáñez, Núria Macià, Víctor Llorente
Universitat d’Andorra (ANDORRA)
Abstract
Over the last few years, the scope of the university, traditionally focused on young adults, has been
revisited; universities have started an opening process towards today’s information-driven society to
contribute to the transformation of our education system. Since 2007, Universitat d’Andorra (UdA) has
embraced this new era by implementing multiple initiatives in the framework of the UNESCO Chair
Information Technologies: Training and solidary development. The case of small states. Among these
initiatives, “Aprèn amb Pitavola” stands out as a contemporary approach to instil technology-related
knowledge into primary schoolers. The main goal of this non-profit initiative is to bridge the gap
between children’s inherent ability to use technology and the actual interest in becoming more than
mere users.
Keywords: STEM, Computer programming, Innovation in education, Primary school, After-school
program
1 INTRODUCTION, CONTEXT AND BACKGROUND
“To reading, writing and arithmetic, we should add
computational thinking to every child’s analytical ability” [1].
It is well-acknowledged that computational science and engineering (CSE) have a key role in the
future of the scientific discovery process and engineering design. However, Glotzer et al. 2009 found
that a worldwide shortage of scientists and engineers trained in the fundamentals of CSE might slow
progress down [2]. Indeed, the labour market has recently seen a concerning shortage in Science,
Technology, Engineering, and Mathematics (STEM) workers, and Andorra has not eluded this
negative trend. In this regard, UdA is trying to revert the generalized steadily decline in Computer
Science (CS) and tech-related vocations by inspiring students to pursue STEM careers.
According to the World Economic Forum Report, students need more than traditional academic
learning. They must be adept at collaboration, communication and problem-solving. Critical thinking,
creativity, communication, and collaboration have been identified as key competencies students
should acquire. The literature suggests that the computing curriculum should include computer
science, information technology, and digital literacy [3] and programming should be introduced at an
early age of schooling [4]. With this in mind, as a first approach, the UdA has designed a
contemporary after-school pilot program for awakening these skills.
Adopting technology in the classroom, which is still an ongoing task, has significantly changed how
contents are taught and learnt. This is currently overlapping with the need for integrating innovation
and technology into the curriculum. Thus, the education field is facing a second revolution unleashed
again by the digital era. A revolution, driven by outsiders to the education system, that takes place at
an unprecedented pace, revolves around topics beyond the qualification of most teachers and school
administrators [4, 5], and challenges the traditionally limited allocation of resources to technology-
related subjects. Indeed, Google highlighted that while a certain part of the society dealing with
education (students, parents, teachers, and administrators) value computer sciences, school
administrators systematically underestimate this demand [6].
Hence, the education system is not meeting the needs of a digital society regards digital skills. New
methodologies within the learning and teaching process are required to substantially modify the
organization model of the educational institutions such as specific training devoted to computational
thinking (CT). CT involves the three A’s key constructs: Algorithms, Abstraction, and Automation [1].
As stated by Yadav, Hong, and Stephenson, "CT involves problems decomposition, using a sequence
of steps (algorithms) to solve problems, reviewing how the solution transfers to similar problems
(abstraction) and finally determining if a computer can help us more efficiently to solve those problems

(automation)" [4]. Thus, CT becomes a transversal skill across fields of study, rather than a
standalone computer science subject. Programing is not just a cognitive skill for coding, but a social
and cultural skill. This is known as “connected learning”: how technology can help in daily problems
[7].
Bringing CT to the classroom, however, is not a new concept. Early initiatives go back to the 80s with
LOGO, the first pedagogical programming language. Seymour Papert designed an environment that
enabled students to learn maths, cybernetics, and science through coding [8]. LOGO is the precursor
of Scratch, one of the most popular platforms adopted in schools lately thanks to its visual interface
and easy-to-use system where users drag and drop instructions. Despite the ease of access to
technology and the availability of learning tools to encourage coding in education, schools still have
been lagging at adjusting their curriculum and updating their resources. The UdA has taken this
challenge up and explored how to implement a program that helps transitioning from tech-based
extracurricular activities to subjects designed to both introduce and hone computational skills at
school. To this end, a multidisciplinary team has examined and evaluated a diverse set of existing
programs, platforms, and activities, and designed a pilot that fits the Andorran needs.
The purpose of this contribution is (1) to describe the ins and outs of the proposed program, (2) share
the experience, and (3) provide an insightful discussion on how to scale nationwide.
The remaining of this paper is organised as follows: Section 2 describes how the program was
designed and implemented and Section 3 discusses the students' learning curves. Sections 4 and 5
concludes with a discussion about the lessons learnt and the essentials for this kind of after-school
program and further work.
2 THE “APREN AMB PITAVOLA” PROGRAM
UdA hosted “Aprèn amb Pitavola” (it literally means Learn with Butterfly in Catalan, official language of
Andorra and uses the Andorran term for butterfly), a non-profit, volunteer-based initiative, that brought
innovation and technology to the Andorran students from primary school.
The pilot launched in September, 2016 with 38 students—ages from 9 to 12—from two primary
schools of different educational systems (Andorran and French; Andorra has three different
educational systems—Andorran, French, and Spanish—all three free of charge and free choice).
During the academic year, students met weekly for one hour and a half at the university to learn how
to code. The curriculum focused primarily on teaching the basics of coding such as variables,
conditionals, loops, algorithm construction, logical reasoning, decomposition, and patterns. The use of
Code Studio provided a fun environment where students could practice new concepts while playing
with computer games and solving puzzles. Part of this pilot was loosely inspired in Girls Who Code
(GWC, https://girlswhocode.com), a movement born in the United States to close the gender gap in
technology. The following sections describe how the program was designed and implemented.
2.1 Design process
In order to launch a program that would inspire youngsters to find a passion in innovation and
technology a four-step process that involved taking decisions on (1) curriculum, (2) tools, (3)
facilitators, and (4) host was designed.
2.1.1 Curriculum: Coding and robotics
CS is not only about learning how to code but learning how to solve problems. This has lately evolved
towards the concept of CT, understood as a structured way to formulate and solve problems based on
computational concepts. Students are now expected to understand how to formulate a problem by
organizing the information logically, be able to decompose the problem by breaking it into smaller
components (or building blocks), implement possible solutions, and generalise the solution to a wider
type of problems [1, 8].
The proposed program has been devised from its foundations to introduce students to the art of
programming and use this tool to develop CT skills. Table 1 shows the core of preliminary
competencies intended to address in the program grouped into three categories: technical,
behavioural, and social.

Table 1. Core competencies of the program
Technical
Behavioural
Social
Basic usage of computers
and their environments
Foreign languages
Tolerance to ambiguity
Face problems using critical
thinking
Confidence and perseverance
when facing complexity
Work autonomously
Motivation to solve challenges
Ability to generate new ideas
Communication skills
Teamwork and
collaboration
We relied on actively engaging students using hands-on activities, one of the best ways for primary
students to learn about CS according to Berry [3].
Besides, the activities have been based on the following principles:
(1) Work in an interdisciplinary way by playing different roles (such as designer or inventor).
(2) Encourage students to work on projects related to their interests.
(3) Build a sense of community among students.
(4) Offer opportunities, resources, and a safe environment based on trust and respect to those
who cannot access digital technology.
2.1.2 Tools for coding, robotics, and additional resources
A. Coding – Code Studio
Nowadays, a wide range of platforms to teach and learn CS are available online. In order to find the
best suited tool for our purpose we narrowed down our analysis to five platforms—Code Studio,
Codesters, Scratch, Python—and assessed aspects such as curriculum, accessibility, pricing, and
support for teachers (also referred to as facilitators).
The following online platforms were evaluated:
Code Studio (https://code.org) is an online education platform that provides teachers with an
integrated environment where students can access level-oriented puzzles and solve them by using
simple drag-and-drop actions—and by doing so, learning the basics of coding. Online exercises and
offline activities are available to seamlessly introduce students of any age to coding concepts and help
them reinforce what has been previously learnt and practised. Teachers have a dedicated webpage to
monitor student progress. Their curriculum is available at no cost.
Scratch (https://scratch.mit.edu) is one of the most well-known platforms to learn how to code and
has been adopted among the schools to initiate students into coding, being the case for some pilots in
Andorra as well. Despite a large community behind, there are no lesson plans available or self-guided
activities. Scratch is more of a learn-by-doing environment where teachers can set up challenges and
students can solve them by using the instruction set available and/or finding support on the community
examples and projects. Unfortunately, no resources were available at the time we designed the
program to develop a full 20-session curriculum. Besides, there is no backend for reviewing student
progress.
Codesters (https://www.codesters.com) is a similar platform to Code Studio and Scratch. There is
an initial free plan to learn how to code. However, the entry level requires advanced fluency in Python.
In this case, students can either drag and drop instructions or directly type them into the console (even
if most of the actual coding is kept behind the curtains). Facilitators can also track student progress.
Codesters extends the courses offered by paying a fee of 10$/student. Additional options were also
available per number of users and a quote was necessary.
Plain Python with no online platform was considered to teach the students how to code straight from
the console. Many books and online resources are available to prepare lesson to teach young kids
Python. Nonetheless, these resources advised that children should get familiar with CS concepts first
using graphical environments like the ones aforementioned.

The chosen platform was Code Studio since it is free, provides many different age-oriented courses
students could start right away, and gives facilitators access to a simple, visual dashboard to follow
students' learning progress. In addition, Code Studio offers countless exercises that do not require
access to devices so that students can still acquire the basics of programming and apply them later on
to any device. As a bonus, students can work collaboratively by sharing a single computer while the
facilitator can still keep track of their improvement.
Code Studio proposes, mainly, two types of exercises: (1) Labyrinths, where students have to use
problem-solving skills to tackle complex situations that require to guide characters through mazes and
similar structures and (2) Artist exercises, where students have to determine the steps a character has
to make or repeat to draw different objects or patterns. Interestingly, the latter regains LOGO's
fundamentals, proving that the methodology implemented forty years ago is still relevant.
B. Robotics – Lego and Parrot drones
We used the following kits to introduce students to robotics and provide a physical experience of the
concepts practised on the computer (mainly because they were resources already available at the
university):
LEGO’s WeDo 2.0 and EV3 platforms combine the LEGO brick philosophy that children have grown
to love over the years with the possibility of adding especially designed controllers and sensors that
allow the built structures to interact with the surrounding environment.
Parrot Jumping Sumo, commonly called floor drone, can be programmed using an iOS or Android
device and allows students to make the drone perform a series of actions. The programming
capabilities were somewhat limited but enough for children to test sequential programming and loops.
C. Additional resources – Hour of Code and TED Talks
Students would also be encouraged to follow other activities to showcase that what they had
previously learnt in Code Studio could be applied to many other aspects of their daily lives.
Hour of Code initiative. One meeting would be dedicated to the Hour of Code. Code Studio supports
one-hour introduction session following the same methodology as the regular courses where students
can solve other puzzles and create a fully working game using known scenarios such as Star Wars,
Frozen, or MineCraft.
TED Talk. One meeting would focus on other coding-related experiences by watching TED talks from
other kids’ experiments.
2.1.3 Facilitators: Volunteer-based
The facilitators volunteered for the program. The team was comprised by teachers from the UdA with
backgrounds in CS and Education, and college students at UdA from the CS degree and the Science
and Education bachelor.
2.1.4 Host
UdA provided the computer labs, that is, two classrooms equipped with 40 desktop computers running
on Microsoft Windows 7 operating system and with Internet access. The students were free to choose
either Mozilla Firefox or Google Chrome as their working browser.
2.2 Implementation
Over the 20 sessions, adjustments were made to provide a better learning environment to the
students. Changes affected collaborative work, working environment, and peer mentoring.
Collaborative work. Initially the program was intended to use a single physical classroom as well as
a single virtual classroom where all students would be registered. This led students to share
resources, that is, one computer per two students and work in a collaborative mode. This required the
student to sign into Code Studio as either a driver or a navigator. Therefore, in the two first meetings,
students worked in pairs (or groups). This was found to be non-productive because all students
wanted to have control over the computer. Also, problems raised when pairing students. At this age,
children want to partner with friends, of the same age, and mostly of the same gender, especially
when new acquaintances have to be made (bear in mind that the class included students from
different schools). Since enough computers were available for each student, soon they started working
with their own computer.

Working environment. Later on, students were separated into two different physical classrooms.
First, this change was motivated by the evidence of two groups of students progressing at a very
different pace. All students started the program on Course 2. In a short period of time, thanks to the
teacher dashboard, we observed that half of the students finished all lessons and had to be moved to
Course 3. Hence, two different virtual classrooms with 20 students each were created. Secondly,
close to 40 children and three/four facilitators discussing assignments in one space was a loud
environment to work in. Having two separate spaces with fewer people notably improved the work
environment.
Peer mentoring. With the aim of not leaving any student behind, students ahead in their lesson plan
who achieved better results were invited to act as peer mentors and help their mates solve their
assignments.
3 LEARNING PROGRESS
Through a diverse set of assignments students became familiar to programing concepts such as
algorithms, logical reasoning, decomposition and pattern finding, and sequential programing, loops,
conditionals, and functions. Students' progress was assessed by tracking attendance, reviewing
assignment completion, and taking a final test.
3.1 Attendance
Fig. 1 shows the attendance records. Thirty-eight students signed up for the program. After discarding
dropouts—eight in total (four boys and four girls)—we can see that the attendance fluctuates with an
average attendance of 25 students. There is a significant drop in the attendance in January 2017 (i.e.,
meetings 6-8). This is due to the fact that students had ski training those days. Even if these two
activities did not share the same timetable, students were, occasionally, too tired to attend the
meetings. Note that students that dropped out the program told us that it did not suit their
expectations.
Figure 1. Attendance records
3.2 Assignment completion
Code Studio teachers' dashboard—using a visual and effective code system—allowed the facilitators
to review students' work and check whether the assignments were not started, in progress, completed
with too many blocks, or perfectly completed.

Students progressed at different speeds. A group of students, without taking age or gender into
account, easily acquired the basic programming skills. These students were able to solve the first set
of puzzles rapidly and completed the Course 2 in few sessions. This introductory course was aimed at
8-year old children and was meant to be completed in 20 meetings. Most students, especially the
older ones, did not need any help from the facilitators and move forward to Course 3. We wondered
whether such rapid completion of assignments also resulted in the assimilation of the corresponding
programming concepts. On the other hand, another group of students required much more time to
solve the exercises from Course 2.
Sequential programming and loops were fully understood, and students were able to identify when
such structures were necessary. On the contrary, conditionals and functions required much more help
from the facilitators so that students could successfully resolve the problems. In this regard, it should
be said that conditionals and functions were introduced much later in the curriculum.
Another result that is worth commenting upon is that throughout many sessions, students were
challenged with applying the learnt CS concepts to their daily routine activities and comparing how
these two were related. In most cases, students acquired skills that also helped them rationalize
common daily problems and activities. When asked, though, if they ever thought of algorithms in their
daily life, their unanimous answer was negative.
3.3 Final test
A written 15-question test (available upon request) was carried out the last day of the program. The
questions were divided into the following five sections:
(1) Multiple choice (text): 7 questions
(2) Multiple choice (images): 3 questions
(3) Write the answer: 1 question
(4) Relation: 1 question
(5) Find the right answer and identify mistakes: 3 questions
Results of the written test show that, overall, the initial goals of "Aprèn amb Pitavola" have been
achieved. However, it is worth noting that many of the questions in the exam were taken straight from
the environment that students used during the program and, even so, some of them failed to answer
them correctly. The fact that it was a written test and the students did not have the possibility of
conducting a trial-and-error approach seems to be the explanation that some of these questions were
answered incorrectly.
After the test, the students were asked about its difficulty. Most of them agreed that it was easier for
them understanding the concepts on the computer than on a piece of paper since this time they could
not apply the trial-and-error methodology to see whether it worked.
In terms of score, there is a significant difference between students in fifth grade (ages 9 to 11) and
students in sixth grade (ages 11 to 12). The first group obtained an average score of 6.29 whereas the
second was of 7.27. Respectively, the medians were 6 and 8, and the standard deviation for each
group was 1.92 and 1.75. The difference in the score, apart from the age, may be that it is not only
until sixth grade that students learn degrees and circumferences. It was found that this type of
exercises was among the ones that students struggled the most.
Not enough data was available to determine whether their origin school had an influence on their
score. Also, no correlation was found between attendance and score. Missing up to 6 meetings was
considered that the student dropped out the program. However, results do not show that students
attending all meetings had better results than those that missed at most 6 meetings.
4 DISCUSSION
The program was offered to two grades in two different schools located near by the university. This
was convenient for the students since the university is at walking distance from both schools. The
advertising of the program was kept to the minimum in order not to overfit the resources available.
Almost 40 students joined the program, significantly exceeding expectations, which were initially set to
15 chairs. Parent Associations from different schools across the country have shown interest in
hosting similar initiatives in coming years. This interest, together with the fact that this was a pilot,
makes it necessary to evaluate how to scale the program nationwide and define a methodology so

that schools and the UdA can work together on progressively introducing the content to the school
curriculum.
Choosing the adequate and suitable tools is considered key to build a successful program and to
properly teach the intended curriculum, manage students' behaviour, and have a devoted team for the
task. In this sense, Code Studio has proven to be a great platform to initiate young kids into coding.
Students' behaviour is a relevant factor in their learning curves. Each student had their own pace and
methodology to solve the proposed problems. Groups of students were formed, and even if each
student had their own computers, they worked collaboratively to solve most of the problems. The
teammate profile for each group was homogenous in terms of age, gender, and school.
Furthermore, regarding behavioural aspects, a few students regarded the program as a sort of after-
school activity more related to having fun and hanging out than a learning activity. This behaviour
made uneasy both facilitators and fellow students and confirms the importance of starting from the
very beginning the purpose of the program and sign a sort of contract among students stating the
values of the program such as respect, hard work, and perseverance. This also leads to consider
having an application process to join the program. Additionally, if behavioural issues or lack of
attendance were detected, students could be expelled from the program.
This initiative was volunteer-based, which is a concept not as extended in Andorra as is in other
countries. When Education and Computer Science college students—and even fellow lecturers—were
asked to be part of the program, a lot of enthusiasm was initially shown. However, the amount of work
and energy required to deal with students from primary school discouraged many volunteers. Although
similar programs recommend a ratio of one facilitator to sixteen students, our experience shows that
for this age, one facilitator for eight students is best. This will help the facilitator keep the class under
control and be able to address all their students doubts and questions.
5 FURTHER WORK
The next challenge is to determine how to successfully scale the program nationwide. To this end, we
need to rethink resource allocation and work on student engagement.
Ideally, building a network of clubs would be the best, where each school in Andorra could host a
"Aprèn amb Pitavola" club. Clubs would follow the same guidelines and curriculum. With such a
network, we would count with a much larger workforce. To this end, we could design a training
program for facilitators; this would help school teachers and volunteers join the movement. For the CS
degree and Science and Education bachelor, becoming a facilitator of one of these clubs could be a
project of one of their subjects.
While transitioning from an after-school activity to a curriculum activity at the school, students can
keep joining the club year after year. Actually, a 70% of students expressed interest in joining next
year. In this case, we should be ready to manage multiple levels and courses simultaneously. One
research line to consider is to work on the Code Studio assignment to learn the core concepts which
are going to be applied to a larger project. This impact project (or a group challenge) can be a drone
race or robot task that students would have to solve working as a team.
Regarding student engagement, a set of bonding activities and icebreakers that help students to hone
soft skills such as not giving up or willingness to participate in class discussions should be developed.
REFERENCES
[1] Wing, J. M. “Computational thinking”. Communications of the ACM, 49(3), 33-35, 2006.
[2] Glotzer, S. C., Kim, S., Cummings, P. T., Deshmukh, A., Head-Gordon, M., Karniadakis, G., &
Shinozuka, M. “International assessment of research and development in simulation-based
engineering and science”. Panel report. World Technology Evaluation Center Inc. Baltimore MD,
2009.
[3] M. Berry. Computing in the national curriculum. A guide for primary teachers. Computing at
school. Educate-Engage-Encourage. Ed. Nace, 2013.
[4] A. Yadav, H. Hong, and C. Stephenson. “Computational Thinking for all: Pedagogical approaches
to Embedding 21st Century Problem Solving in K-12 Classrooms”. TechTrends 2016.

[5] CSTA. “Bugs in the system: Computer science teacher certification in the US”. New York, NY:
Computer Science Teachers Association, 2013.
[6] Google. “Searching for computer science. Acces and barriers in U.S. K-12 education”, 2015.
[7] J. Valverde, M. R. Fernández, and M. C. Garrido. “El pensamiento computacional y las nuevas
ecologías del aprendizaje”. RED-Revista de Educación a distancia, 46(3), 2015.
[8] S. Papert. Mindstorms: Children, computers, and powerful ideas. Basic Books, Inc., 1980.